ni_mio_common: Bug: DMA overwrite of free Area

Hallo

While I find the error shown in the subject in my kernel log, my problem 
is most probably related to my incomplete understanding how the commands 
work for the streaming input.

I just started using comedi 0.7.66 and comedilib  0.7.19 on a Suse 8.1 
System with a NI PCI 6036E. I develop with Labview 6.1 using the VI's 
from the NI -site, which are basically wrappers for the corresponding 
comedi functions. DIO works fine, but the AI gives me some grief:
I first send a comedi_command with stop_source none and stop_argument 
0 . Then I start a for-loop which reads in 4080 data points in each 
iteration. The sample frequence is 1 kHz and the loop needs longer than 
5 s, so there should be enough new data each time. This works for the 
first 3 iterations fine, the fourth contains some garbage in the first 
250 data points and at some point later the program freezes. Moreover I 
get the above error message.

I thought the data are inserted in a kind of ring buffer, and as long as I 
don't want to read chunks that are bigger than the buffer size (where 
can I find that number ?) I don't have to worry about the actual position 
of the pointer. 
The only reference to my error message I found in the archive was:
https://cvs.comedi.org/pipermail/comedi/2002-October/003655.html
but this seems to be fixed, at least the patch won't apply any more.
 
Many thanks for your attention.

                        Matthias


##############################################################################
Matthias Schröter

Center for Nonlinear Dynamics
University of Texas at Austin
1 University Station C1610
Austin, Texas 78712

Phone: ++ 1 512 471 5926
Fax:   ++ 1 512 471 1558
############################################################################## 

Received on 2003-05-22Z01:48:10